External Microflora of Marine Wood-Boring Isopod · Wood-boring isopods are ubiquitous in tem-perate and tropical marine habitats. They are responsible for substantial damage to equip-ment,

Vol. 42, No.4APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Oct. 1981, p. 720-7290099-2240/81/100720-10$02.00/0

External Microflora of a Marine Wood-Boring IsopodPAUL J. BOYLE* AND RALPH MITCHELL

Laboratory ofMicrobial Ecology, Division ofApplied Sciences, Harvard University,Cambridge, Massachusetts 02138

Received 19 January 1981/Accepted 16 July 1981

Bacteria associated with the marine wood-boring isopod Limnoria lignorumwere enumerated by acridine orange epifluorescence microscopy and by platecounts on several media; the plate-viable bacteria were isolated and identified.Similar procedures were followed to enumerate and identify bacteria associatedwith the wood substrate from which the isopods were collected and with thesurrounding water from the isopod habitat. Approximately 1.4 x 107 bacterialcells were associated with each individual L. lignorum. Aeromonas hydrophila,Pseudomonas, and Vibrio were the most common genera in the isopod microflora.Wood from L. lignorum burrows had an associated bacterial flora of 1.6 x 107cells per mg (damp weight). A. hydrophila also predominated in the woodmicroflora. The water from which the isopod population was collected contained2.3 x 106 bacteria per ml. Pseudomonas and Vibrio species were very common inthe water microflora, but A. hydrophila was not detected. Interactions betweenthe isopod, its associated microorganisms, and the microorganisms within thewood substrate are discussed in the light of the known absence of a residentdigestive tract microflora in these animals.

Wood-boring isopods are ubiquitous in tem-perate and tropical marine habitats. They areresponsible for substantial damage to equip-ment, structures, and boats in estuarine andcoastal zones. Wood biodeterioration costs theU.S. Navy more than $200 million annually (20),and private losses are also large.

Species ofLimnoria maintain a digestive tractwhich normally is free of any resident, attachedmicroflora (6, 7, 45). An exception to the normalbacteria-free gut condition occurs in L. tripunc-tata removed from creosoted wood (58); bacte-ria, representative of the microflora found onaged creosote, were found within the digestivetract of these isopods.The complete lack of a microflora in the nor-

mal Limnoria digestive system is unique in lightof present assumptions concerning the microbi-ological status of animal digestive tracts (10).Conversely, the outer exoskeleton surfaces ofthese wood-borers are colonized densely by adiverse array of bacteria and other microorga-nisms (6, 7).There have been few investigations of the

microflora associated with marine invertebrates(52). This report contains the results of a studyof the respective bacterial floras found in asso-ciation with the exoskeleton of Limnoria ligno-rum (Rathke), with the wood from which theisopods were collected, and with the seawater inwhich they were found. The objective of this

investigation was to determine whether a char-acteristic bacterial flora is associated with theseisopods or with their wood substrate.

MATERIALS AND METHODSSample collection. Two samples of untreated pine

(20 by 5 by 2 cm) were removed from the wood frameof a specimen tank which had been submerged for 7months beneath the Marine Biological Laboratorydock in Eel Pond, Woods Hole, Mass. (salinity, 29%o).The wood contained a very active population of L.lignorum (Rathke), as well as the wood-inhabitingamphipod Chelura terebrans Philippi. No macro-scopic fouling was present on the wood surface. Alco-hol-rinsed (95% ethanol) latex gloves were worn duringthe collection of all samples. The samples were placedimmediately in a polyethylene pail (prerinsed with95% ethanol) filled with Eel Pond seawater. No at-tempt was made to exclude microorganisms from theair-water surface microlayer, since the wood contain-ing the Limnoria population had been exposed contin-ually to this microlayer while in situ.Water samples were collected from the same site.

Sterile Whirl-Pack bags (approximately 14 by 20 cm)were immersed in the water and held vertically 15 cmbelow the surface. While submerged, the perforatedtop was ripped from the bag and the wire mouth wasopened, thus allowing the entrance of a water sampleof approximately 400 ml. The bag was raised out ofthe water with its mouth still opened to include micro-organisms from the surface microlayer. The bags weresealed with their integral wire mouths immediatelyupon removal from the water and were placed in a

720

VOL. 42, 1981

large bucket of seawater to provide temperature con-trol during the short trip to the laboratory. Analysisof the respective microfloras began within 30 min ofLimnoria, wood, and water collection.

Bacterial isolation and culture media. Foursolid media were utilized for the initial isolation andenumeration of plate-viable, heterotrophic aerobicbacteria. Except for the first (listed below), the mediawere made in artificial seawater, which consisted ofNaCl, 24.0 g; MgCl2.6H20, 5.3 g; MgS04.7H20, 7.0 g;KCl, 0.7 g; and glass distilled water, 1 liter.The media were: (i) marine 2216 agar (Difco Labo-

ratories, Detroit, Mich.) (2216 agar); (ii) marine stan-dard methods agar (BBL Microbiology Systems, Cock-eysville, Md.) made with artificial seawater (MSMagar); (iii) one-tenth strength MSM agar made withfull-strength artfficial seawater (MSM-1 agar); and (iv)natural seawater agar consisting of aged (2 months),unfiltered seawater from Nahant, Mass., plus agar(NSW agar). All media were sterilized by autoclavingand contained a final concentration of 1.5% agar.

After the initial isolation, bacteria were purified andmaintained on plates and slants on 2216 agar incu-bated at room temperature (18 to 22°C).Specimen preparation and bacterial plating.

All dissection equipment, tissue grinders, and petridishes were sterile; all fluids were sterilized by filtra-tion (0.22-,um sterile Millex filters [Millipore Corp.,New Bedford, Mass.]). Specimens of L. lignorum wereremoved from the wood by dissection under a Wild M-8 zoom stereomicroscope. The animals were placed in20 ml of sterile seawater in a petri dish. The animalswere allowed to swim in the filtered seawater for 10min to remove bacteria that were not attached per-manently to the exoskeleton.

After the sterile rinse, single L. lignorum individuals(1.5 to 3 mm, approximate length) were placed in eachof 10 conical glass tissue grinders (3 ml nominal vol-ume; Bellco Glass, Vineland, N.J.) containing 2.0 mlof filtered sterilized NSW collected at Woods Hole.The animals were homogenized for 2 min (clearancein the cylindrical portion of the grinder was 102 to 152,um). Samples were ground slowly with frequent 5-srests to prevent excess frictional heating of the sample.After homogenization, three serial 10-fold dilutions ofeach L. lignorum homogenate were made, using sterileartificial seawater with 0.1% (wt/vol) peptone as dilu-tion water. A 0.5-ml subsample of the homogenateremaining in each tissue grinder was transferred to aseparate glass vial containing 4.25 ml of filtered (0.22,um) seawater and 0.25 ml of filtered (0.22 ,um) 40%Formalin. These preserved samples were retained at5C for epifluorescence direct counting of bacteria.

Samples (0.1 ml) of each homogenate and its re-

spective 10-1, 10-2, and 10-3 dilutions were inoculatedindividually onto duplicate plates of the four agarmedia (Fig. 1). Inocula were spread over the agarsurface with an alcohol-flamed glass rod, using a stan-dard petri dish turntable. After inoculation, plateswere covered with their sterile tops, inverted, andincubated at 20°C.

During the aseptic collection of L. lignorum, threesmall samples of the heavily bored pine habitat wereremoved with a sterile scalpel and forceps. Each sam-

ple was taken from within a separate isopod burrow,

WOOD-BORING ISOPOD MICROFLORA 721

but no L. lignorum individuals were included. Thethree wood samples were weighed in separate, sterile,preweighed aluminum foil envelopes. After weighing,the wood samples were homogenized in 2.0 ml of filter-sterilized NSW, as described for L. lignorum individ-uals. A 0.5-ml sample of each homogenate was pre-served for acridine orange direct counting (AODC) asdescribed. Samples (0.1 ml) of each 10-2, 10-3, and 10-4dilutions were spread plated in duplicate on each ofthe four media, as described for isopod homogenates(Fig. 1).The water samples were shaken to distribute the

contents evenly within the Whirl-Pack bags. A 1.0-mlsample was removed aseptically from each of foursampling bags with separate sterile 1.0-ml pipettes.Three serial 10-fold dilutions were prepared from eachwater sample. Samples (0.1 ml) of the 10' to the 10-3dilutions were spread plated in duplicate on each ofthe four media (Fig. 1). A 0.5-ml sample of each 10-'dilution was preserved for AODC, as described.

Plates were stored in sealed plastic bags duringincubation to prevent drying. Colonies were countedon the appropriate plates after a 3-day incubation andagain after 10 days. The NSW plates were incubatedfor 1 month before colony counts were made to insurethat all plate-viable cells had formed colonies. Colonieson the NSW plates were extremely small, and evenafter 30 days of incubation counting required the aidof a Wild-M8 zoom dissecting microscope at approxi-mately x40 magnification. The NSW plates were uti-lized for enumeration purposes only; no isolates werecollected from these plates.

Epifluorescence enumeration. Preserved sam-ples for AODC were stored at 5°C and were analyzedwithin 1 month. Methods for AODC were the same asthose developed by Hobbie et al. (26).

Counts of AO-stained cells were made with anOlympus Vanox research microscope equipped with a200-W mercury burner, a fluorescein isothiocyanateexciter filter, and a BG-12 exciter filter, a DM-500dichroic mirror, an 0-515 filter, and a Y495 barrierfilter. Fields for counting were obtained by randommovement of the microscope stage from its initial,near-central location; stage movements were carriedout while the operator's eyes were closed, to eliminatebias in the choosing of fields for counting. Two filterswere prepared for each sample, and twenty fields werecounted on each filter.

Selection and identification of bacterial iso-lates. Bacterial isolates were collected from all indi-vidual L. lignorum, gathered from widely separatedburrows in two different wood samples, to provideinformation representative of the L. lignorum popu-lation as a whole.The large number of colonies that grew on the

spread plates precluded the isolation of every colonyfor identification. Since many bacteria have nearlyidentical colony appearance on agar plates, isolates foridentification were selected as follows. For the first 2of the 10 L. lignorum individuals that were sampled,virtually all of the bacterial colonies on the countableplates were isolated and purified for identification,including many that looked identical, to insure thatnone were overlooked simply because they had themorphology and color of another colony. Several col-

722 BOYLE AND MITCHELL

Sample Type

1. individualL. lignorum

or

2. individualwood somplefrom burrow

Tissue grinder

0.5 ml !

AODCSample

*GNSW C2216 CMSM CMSM-'C

or 3. individualwatersamplel.Oml

l.Oml ,1.Oml

10- 1

0.lml

)O 00)O 00)O 00)O 00

.OmIm 1.0m1

-13 -c~410 10-(Sampletype 2

v .! t only)00 0000 0000 0000 00

*-4.25ml 0.22pL filtered seawater-0.25ml 0.22/i filtered 40% formalin

FIG. 1. Sampling of bacterial microflora associated with the wood-boring isopod L. lignorum, wood fromL. lignorum burrows, and water from the L. lignorum habitat. Numbers within tubes indicate volume ofhomogenization or dilution fluid.

ony types had a very distinctive morphology and colorpattern, and for these, obvious duplicates were notisolated repeatedly. Additional bacterial isolates werecollected with the intent of gathering the full range ofplate-viable bacterial types present. Consequently, theplates from the remaining eight L. lignorum weresearched for obviously different colony types, andthese were isolated and purified for identification.

Bacterial isolates from the wood and water sampleplates were collected by the same methods outlinedfor isolate collection from L. lignorum plates. That is,almost all colonies were isolated from the appropriateplates of the first sample in a given series regardless ofapparent duplication (e.g., almost all colonies from thefirst wood sample and the first water sample); platesfrom the remaining samples in a given series served toprovide isolates representative of the range of plate-viable bacteria associated with the type of sampleunder study.

Isolates were collected on 2216 agar from distinct,well-separated colonies. Once isolated, bacteria werepurified by streaking on 2216 agar plates. Cultureswere Gram stained and observed microscopically fordetermination of cell morphology and cell arrange-ment. It was possible to place gram-positive isolates in

a genus or major group with only a few additionaltests. These included cell morphology parameters andtests for motility, pigment production, production ofacid or gas when grown on glucose, the presence ofcytochrome oxidase and catalase activity, and sporeproduction. Gram-negative isolates were identifiedwith the use of: (i) the API 20E system (AnalytabProducts, Plainview, N.Y.); (ii) Oxiferm tubes (RocheDiagnostic Laboratories, Nutley, N.J.); and (iii) addi-tional, standard bacteriological tests not included in,but in some cases duplicating, the previously men-tioned systems (8, 16, 33). Additional, peripheral testsincluded oxidase and catalase activity, motility, salttolerance and requirement, growth on MacConkeyagar, nitrate reduction, denitrification, agar liquefac-tion, utilization of lactose, pigment production, andduplication of tests contained in the prepackaged sys-tems using laboratory-prepared media. Isolates thatexhibited a salt requirement were suspended in 3.0%NaCl, rather than 0.85% NaCl, for inoculation into theAPI 20E system.

Comparison of the biochemical and morphologicalcharacteristics of the isolates with data and identifi-cation schemes contained in Bergey's Manual of De-terminative Bacteriology (8), the Color Atlas and

APPL. ENVIRON. MICROBIOL.


Textbook of Diagnostic Microbiology (34), and SeaMicrobes (49) allowed identification to the genericlevel for most isolates and to the species level formany. Several additional references aided in the iden-tification of the bacterial isolates (2-4, 11, 14, 21, 32,35, 47).

RESULTS

Bacteria associated with the isopod exoskele-ton, the wood burrows, and the surroundingwater were enumerated, isolated, and identifiedto allow a comparison ofthe plate-viable fractionof the respective microfloras.Bacterial enumerations. The 2216 agar and

the NSW agar gave nearly identical counts forall samples (Table 1). In all cases, both MSMand MSM-1 agars gave counts that were 10 to100 times lower than those obtained with 2216or NSW agar. The MSM agar and the MSM-1agar had the additional drawback of drasticallyreduced colony diversity when compared withthe NSW and the 2216 media. MSM and MSM-1plates routinely showed only one or two colonytypes. For the reasons of lower numbers andreduced diversity, no isolates were collectedfrom the MSM and MSM-1 plates. (The very

few colony types seen on these plates also oc-

curred on 2216 agar, and they were identifiedfrom that medium.)The mean number of plate-viable bacteria

associated with an individual L. lignorum was

1.7 x 105 (±0.9 X 105; 95% confidence level[confi) on 2216 agar and 1.5 x 105 (+0.6 x 105;95% conf) on NSW agar. These means share thesame population variance (Fo.5(9,9) = 4.03; Fsampie= 2.10), and in all cases (Table 1) there was no

significant difference (95% conf) between themeans for 2216 agar and for NSW agar (Studentt tests and the Wilcoxon nonparametric signedrank test). Conversely, counts obtained withMSM and MSM-1 media did not share the samepopulation variance with 2216 agar (F test), andBehrens-Fisher tests (for samples with differentpopulation variance) indicated that the mean for2216 agar was significantly different (i.e., larger)from the means for MSM and MSM-1 agar (51);the significance of this difference was confirmedat the 1% level by the Wilcoxon nonparametricsigned rank test (51).

Previous research (9) had indicated thathigher counts might result with the NSW me-

dium. However, the 2216 agar yielded countsthat were statistically equivalent to those ob-tained with NSW agar. Colony development on

NSW agar took over 1 month, and many colonieswere extremely small, necessitating the use of adissecting microscope for counting. All bacterialisolates for identification were collected from the2216 agar plates. Isolates were not collected from

TABLE 1. Enumeration of bacteria in L. lignorumindividuals, wood samples, and water

No. of bacteria (CFU)" in:

Medium' . Wood' from Water from L.Individual L. L. lignorum lignorum hab-lignorum burrow (per itat (per ml of

mg of wood) water)

2216 1.7 x 105 2.4 x 105 5.4 x 104(1.23 x 105) (0.42 x 105) (1.45 x 104)

NSW 1.5 x 105 2.1 x 105 2.0 x 104(0.85 x 105) (0.32 x 105) (1.16 x 104)

MSM 1.2 x 103 9.4 x 103 5.6 x 103(0.51 x 103) (0.72 x 103) (2.75 x 103)

MSM-' 1.2 x 103 3.3 x 103 4.4 x 103(0.67 x 103) (2.26 x 103) (1.03 x 103)

aCFU, Colony-forming units. Counts shown aremeans (from L. lignorum, n = 10; L. lignorum wood,n = 3; L. lignorum habitat water, n = 4) of duplicateplate count (n = 2) means. Values in parentheses arestandard deviation.

'See text, Bacterial isolation and culture media.'Wood sample actual weights were 8.7, 7.3, and 8.6

mg.

NSW agar since the colonies on these plateswere over 1 month old when they were ready forcounting.Samples of seawater collected from the L.

lignorum habitat contained 5.4 x 104 (±2.3 x104; 95% conf) plate-viable bacteria per ml on2216 agar (Table 1).Samples of wood from L. lignorum burrows

contained 2.4 x 105 (±1.0 x 105; 95% conf) plate-viable bacteria per mg on 2216 agar (Table 1).AODC was carried out on a limited basis due

to problems of bright background fluorescenceof wood and animal tissue. AODC enumerationswere conducted on two of the L. lignorum sam-ples, two of the water samples, and two of thewood samples (Table 2). It is interesting, but notsurprising, that in each case the AODC enumer-ation indicated a bacterial density approxi-mately two orders ofmagnitude higher than thatobtained by plate counts.The enumeration data are summarized in Ta-

ble 2, together with a normalization of bacterialdensity per unit weight for each sample type.Although this normalization is a somewhat ar-tificial manipulation of the enumeration data, itprovides a comparison of the relative bacterialdensities of the various samples.Bacterial identification. We compared the

predominant heterotrophic, aerobic bacteria liv-ing on the exoskeleton of L. lignorum with thespecies distribution in the wood habitat and thesurrounding water. The great majority of bac-teria isolated from all samples and specimens inthis study were gram-negative rods. Four genericgroups comprised the predominant taxa, includ-

VOL. 42, 1981


TABLE 2. Bacterial enumeration summary andnormalized bacterial densities of the bacterial

floras associated with L. lignorum, wood samples,and water samples

Bacteria (CFU) per Bacteria (CFU) perindividual sample' g of sampleb deter-

Sample type determined by: mined by:

Plate AODCd Plate AODCcounte count

L. lignorum (x= 1.7 x 105 1.4 x 107 1.1 x 109 9.3 x 10100.15 mg/isopod)Y

Wood from L. lig- 2.4 x 105 1.6 x 107 2.4 x 108 1.6 x 1010norum burrows(sample = 1mg)

Eel Pond water 5.4 x 104 2.3 x 106 5.4 x 104 2.3 x 106(1-ml sample)

'Data summarized from Table 1 and text. CFU, Colony-forming units.

bNormalization per gram of sample calculated from respec-tive counts obtained by plate count and AODC.

' Difco 2216 agar.dCounts shown are means (n = 2) of individual sample

means (n = 20).en = 26.

ing Aeromonas, Pseudomonas, Vibrio, and Aci-netobacter. However, these genera did not occurwith equal frequency in all of the sampled mi-crofloras.Aeromonas hydrophila predominated in the

microflora associated with L. lignorum and withwood from L. lignorum burrows (Table 3). How-ever, A. hydrophila was not found in any of thewater samples collected from the L. lignorumhabitat in Eel Pond (Table 3).Pseudomonas species were numerous in all

samples, and they predominated in the seawatersamples collected from the L. lignorum habitat.P. fluorescens appears to be the most commonspecies in this group, followed by P. paucimo-bilis, P. stutzeri, P. cepacia, P. putrefaciens,and 15 as yet identified (and not necessarilydifferent) Pseudomonas species.

Vibrio species also occurred commonly in allof the sampled microfloras. V. alginolyticus wasfound in all samples.Acinetobacter species occurred in all samples.

A. calcoaceticus var. anitratus occurred only inthe wood microflora and was a very common

TABLE 3. Bacteria associated with L. lignorum exoskeleton, wood from L. lignorum burrows, and waterfrom the L. lignorum habitata

Relativefrequency Bacteria associated with L. lignoruma Bacteria associated with wooda Bacteria associated with wateraof occur- exoskeleton from L. lignorum burrow from L. lignorum habitatrenceb

M.P.O. A. hydrophila (8)C A. hydrophila (1) Pseudomonas species (4), in-cluding P. fluorescens (1),P. paucimobilis (2), uni-dentified spp. (1)

V.C.O. Pseudomonas species (12), in- A. calcoaceticus var. ani- Vibrio sp. (1)cluding P. fluorescens (3), P. tratus (1)paucimobilis (2), P. stutzeri(1), P. diminuta (1), unidenti-fied spp. (5)

C.O. V. alginolyticus (3), Aerococcus V. alginolyticus (1), Pseu- F. halmephilum (1)spp. (6) domonas spp. (3), Bacil-

lus sp. (2)

Also A. calcoaceticus var. lwoffi (2), Arthrobacter sp. (4), Mi- A. calcoaceticus var. lwoffiErwinia sp. (1), Micrococcus crococcus sp. (1), Plano- (1), C. diffluens (1), Plano-sp. (2), Planococcus sp. (1), coccus sp. (1), Actino- coccus sp. (1), Bacillus sp.Arthrobacter sp. (2), Alcali- myces sp. (1) (1), Actinomyces sp. (1)genes sp. (2), CDC groupIV-C Alcaligenes-like (1),dHyphomicrobium-like species(2), Actinomyces sp. (2)

a Colected at Eel Pond, Woods Hole, Mass., habitat of L. lignorum.b Relative frequency of occurrence on 2216 agar plates: M.P.O. = most predominant organism (40 to 50% of

total number of isolates); V.C.O. = very common organism (30 to 40% of total number of isolates); C.O. =common organism (10 to 20% of total number of isolates); also = one or two isolates only.'Numbers in parentheses indicate the number of identified isolates that belonged to a particular generic

grouping or unidentified class.d CDC = Centers for Disease Control, Atlanta, Ga.



organism in this habitat. A. calcoaceticus is theonly species designation within this genus. Twosubspecies are recognized presently, A. calcoac-eticus var. Iwoffi, which is non-saccharolytic andis recognized as being the more common strain,and A. calcoaceticus var. anitratus, which isoxidatively saccharolytic on glucose (34). A. cal-coaceticus var. Iwoffi occurred infrequently inassociation with L. lignorum and in water sam-

ples from the L. lignorum habitat, but was notfound in specimens of wood.

Gram-positive bacteria were isolated from all

specimens sampled in this study. The gram-pos-

itive organisms were much less common in watersamples from the L. lignorum habitat than inthe microfloras associated with the isopods andthe burrow wood. Aerococcus species were com-

mon in the L. lignorum flora. Other gram-posi-tive cocci, including Micrococcus spp. and Plan-ococcus sp., occurred also in association with L.lignorum and with the wood from L. lignorumburrows. Gram-positive coryneform bacteriawere isolated from the isopod and wood micro-floras where they occurred with low to moderatefrequency. These isolates were assigned to thegenus Arthrobacter on the basis of their beinggram-positive, pleomorphic, nonfermentativerods. Arthrobacter species are among the dom-inant bacteria in soils (8), and their occurrence

in a shallow tidal pond, like Eel Pond in WoodsHole, is not surprising. Gram-positive rods in thegenus Bacillus were common in the wood mi-croflora and less common in the water samples.Bacillus species were not found in the isopodmicroflora.

Additional bacteria occurring in low numbersincluded the gram-negative species Erwinia andAlcaligenes, in association with L. lignorum; thestrongly agarolytic Cytophaga diffluens in watersamples from the L. lignorum habitat; and Fla-vobacterium halmephilum, which occurredcommonly in water samples. Two isolates of a

distinctive, prosthecate bacterium were col-lected from the L. lignorum microflora. Theseisolates were assigned to the genus Hyphomicro-bium on the basis of their very characteristicmorphology (8).

Actinomycetes and yeasts occurred frequentlyin association with all of the samples collectedfor this study.The major bacterial groups associated with

each sample type are summarized in Table 3.

DISCUSSIONThe quantitative and qualitative study of the

surface bacterial population of L. lignorum andits wood substrate was designed to provide an

insight into the comparative microbial ecology

of these closely associated microbial habitats.The bacterial enumerations in this study were

conducted to provide information regarding thetotal microflora associated with L. lignorum andits wood substrate. The bacterial identifications,carried to the generic and, where possible, thespecies level, were carried out to provide a tax-onomic and nutritional description of the plata-ble portion of the total aerobic, heterotrophicbacterial population. These identifications wereintended as a basis for comparison between theanimal, wood, and water microfloras, to deter-mine whether these wood-boring isopods possessa specific flora which is different from that ofthe surrounding environment, with the realiza-tion that the identified bacteria are thoseadapted for growth on plates.The single most noticeable finding in this

study was the clear predominance in the isopodand wood microfloras of A. hydrophila, an or-ganism which was not found in any of the watersamples from the isopod habitat.A report on the occurrence of A. hydrophila

in limnetic environments (48) suggested that thisspecies may not be indigenous to the marineenvironment. Another recent study (31) showedthat, although A. hydrophila occurs frequentlyin low-salinity estuarine waters, numbers of A.hydrophila were consistently low in waters witha salinity above 15%o. Data collected in the pres-ent study suggest that, although A. hydrophilamay not occur commonly in marine water sam-ples, it does enter into close association withmarine wood-boring isopods.The predominance of A. hydrophila in the

microflora associated with wood-boring isopodsand their burrows is a clear departure from thedata of previous studies, in which Aeromonasspecies have been found to occur in relativelylow numbers in association with invertebratescompared to Pseudomonas and Vibrio species(30, 52).

Interest in A. hydrophila has increased duringthe past few years since this potential pathogen(17) has been isolated from uncontaminatedfreshwater supplies, well water (36), and seawa-ter (34), as well as sewage-contaminated water(8). In humans, A. hydrophila can cause tran-sient diarrhea, wound infections, and rarely, sep-ticemia (34), as well as meningitis, gastroenteri-tis, and peritonitis (17). This species is also a

pathogen of snakes and causes red leg disease infrogs and infections in freshwater fish (8, 34),including salmon (49). However, A. hydrophilais not always a pathogen. The gut flora of sal-monids is predominated by Aeromonas spp.when the fish are in freshwater (57), and in clamlarval cultures (species unnamed) sickness can

be associated with a drop in numbers of Aero-

VOL. 42, 1981


monas and other bacterial species (36). Kanekoet al. (30) showed that A. hydrophila was non-pathogenic to the soft-shell clam Mya arenaria.Aeromonas sp. occurs in the gut of the seaurchin Echinus esculentus under normal cir-cumstances (52), and A. hydrophila is associatedwith the freshwater green hydra (Hydra viridis;5). Bergey's Manual (8) states that Aeromonasproteolytica (now designated as a subspecies ofA. hydrophila [8]) was isolated originally fromthe alimentary tract of L. tripunctata (39, 40).This original isolate of A. proteolytica [sic] wasmost probably a contaminant from the exoskel-eton microflora, since Limnoria species now areknown to maintain a digestive tract that is freeof bacterial colonization (6, 7, 45, 46).

In contrast to the absence of a resident micro-flora within the digestive tract, the marine wood-boring isopods possess a dense microflora ontheir external exoskeleton surface (6, 7). At-tached microorganisms are particularly numer-ous on the posterior appendages, the pleopods(6, 7), which function in swimming, respiration,and excretion (43) (Fig. 2).The bacteria found in association with L. lig-

norum quite possibly benefit in several waysfrom attachment to the pleopod surfaces, whereexoskeleton bacterial density is highest (Fig. 2).These bacteria may utilize waste products re-leased by the animal and water-soluble mono-and disaccharides released through the action ofthe cellulolytic flora found within the burrowsystem. Furthermore, A. hydrophila may be welladapted to maintaining a predominant positionin the L. lignorum microflora, owing to its abilityto produce enzymes that lyse a variety of otherbacteria (1).Ammonium is the major nitrogenous waste

product of marine, freshwater, and terrestrialisopods (18, 25, 33, 54-56). These isopods releasea large portion of their nitrogenous wastes acrossthe respiratory surfaces (54). Bacteria attachedto the respiratory appendages may utilize isopodwastes transported across the pleopod surfaces.The large number of bacteria associated with L.lignorum (Table 1), the predominance of A.hydrophila in the L. lignorum microflora (Table3), and the especially dense colonization of rod-shaped bacteria on the pleopods of L. lignorum(6, 7, 50, and Fig. 2) together suggest that A.hydrophila occurs in large numbers on the sur-

face of the pleopods. This suggestion is sup-

ported by the knowledge that A. hydrophila isable to utilize ammonium as a sole nitrogensource (8, 40). The association of A. hydrophilawith L. lignorum pleopods is currently beingstudied in our laboratory by fluorescent-anti-body techniques.Limnoria individuals groom their appendages

by passing them over the mouth parts (53);grooming of the pleopods may result in theingestion of bacteria. The dense external micro-flora (Fig. 2) may represent a type of mixed,continuous culture that is provided with a sur-face for attachment and is supported by theanimal wastes, while it is maintained in an ac-tively growing state and prevented from over-population by the isopod grooming activities.Such a relationship could provide a conserva-tive, closed-loop recycling of at least a portion ofthe dietary nitrogen, allowing subsistence onrelatively nitrogen-poor food materials, such asthe wood on which Limnoria feed. Experimentsare in progress to determine whether such arecycling via the external microflora occurs.Other bacteria that were common in the iso-

pod and wood microfloras include Vibrio, Pseu-domonas, and Acinetobacter species (Table 3),and these have been shown to hydrolyze cellu-lose (1, 8, 19, 27, 28, 44, 49). In addition to theircellulolytic capabilities, some Vibrio strains areknown to hydrolyze chitin, the main componentof isopod and other crustacean exoskeletons (11,49).

Vibrio species have been isolated from a widerange of marine samples. Members of this genuswere common associates of L. lignorum, andthey occurred frequently in the isopod burrows,as well as in the water samples. Pseudomonasspecies and Vibrio species were the most com-mon bacteria associated with the Pacific oyster,Crassostra gigas (12, 13), and similar findingswere reported for the American oyster, C. vir-ginica (38, 41, 42). Similar data were reported ina study of the bacterial associates of flatfish (37).Vibrio species are known to predominate insalmonid digestive tracts when the fish are inseawater (57), and Vibrio and Pseudomonasspecies are common gut inhabitants in mullet(22-24). In addition, Vibrio spp. have been foundin the hemolymph of the blue crab, Callinectessapidus (15), and are known to occur commonlyin seawater (2), in association with copepods(29), and in coral mucus (19).The presence of A. calcoaceticus var. anitra-

tus (capable of oxidative utilization of glucose)among the common wood bacteria suggests thepresence of soluble cellulose breakdown prod-ucts within the wood matrix, possibly releasedby bacterial and fungal extracellular cellulolyticenzyme action and available to the microbialcommunity in general and, perhaps, the boringisopods. It is interesting that A. hydrophila,Pseudomonas spp., Vibrio spp., and Acineto-bacter spp., the four most common genera foundin association with the marine wood-boring iso-pods and their burrows, all grew on MacConkeyagar, which contains surface-active agents that



FIG. 2. Scanning electron microscope view of the dense bacterial colonization typically found on theexternal exoskeleton surfaces of the posterior respiratory appendages (pleopods) of L. lignorum. (A) Low-magnification view showing the hair-covered spines that project from the edge of a pleopod blade and manyattached bacteria (scale bar = 5 pn). (B) High-magnification view of bacteria on the surface of a pleopodblade, showing a diversity of cell morphologies and clear evidence of the extracellular polymer that aids inbacterial attachment (scale bar = 1 um).

inhibit the growth of many bacteria. The isopoddigestive tract contains surface-active agents(unpublished data), and the release of thesesubstances in fecal material may apply selectionpressure for a microflora able to cope with sur-

factants.

The relatively low number of gram-positivebacteria found in all samples of this study is notsurprising. The great majority of bacteria iso-lated from marine habitats are gram negative (4,49). The exceptions to this rule are found insediments and the marine surface microlayer,

VOL. 42, 1981


where relatively high concentrations of organicmatter occur.

It appears from data collected in this studythat 2216 agar is a good medium for enumeratingand isolating plate-viable marine, aerobic, het-erotrophic bacteria. Bacterial numbers detectedwith 2216 agar were the same as with NSW agar.However, colonies grew much more rapidly on2216 agar, facilitating the isolation of viable bac-teria. Bacteriological agar contains, in additionto the cold-water-insoluble neutral galactans,numerous soluble polysaccharides, including py-ruvated galactans and sulfated galactans, whichmarine bacteria reportedly utilize easily (49). Itis surprising that colony formation on the NSWmedium continued for over 1 month, duringwhich time new colonies appeared regularly andgrew very slowly. Although it contained unde-fined constituents, the NSW agar obviously pre-sented a very different nutrient level than the2216 agar. Apparently, the NSW agar containedeither a very low concentration of nutrients or aset of organic contaminants that were not uti-lized readily by the bacteria under study. Fur-ther research is needed to determine whether ornot 2216 agar yields the same diversity of bac-terial types as NSW agar and other media.The number of viable cells in all samples was

approximately two orders of magnitude lowerthan the number of total cells detected by theAODC method. This 100-fold difference be-tween total cell count and viable cell count iscommon in microbiological enumeration studies.Research on media modification may allow thelaboratory growth of a greater proportion of thetotal flora present in sampled materials.Limnoria boring activity greatly increases the

surface area within the wood matrix. The exten-sive surface area within the burrows is madeavailable to colonization by cellulolytic and non-cellulolytic bacterial and fungal species. Thecomplex network of the burrow system exca-vated by Limnoria species increases the availa-bility of native cellulose to the burrow micro-flora. This flora is, in turn, available for grazingby Limnoria species as they move within theburrow system. The microflora of the burrowmay represent an important source of availablenitrogen for Limnoria spp. as they feed on acarbon-rich substrate that contains very littlenative nitrogen.

ACKNOWLEDGMENTSThis research was supported by Office of Naval Research

Oceanic Biology Program contract no. N00014-76-C-0042.We thank Ruth D. Turner and James J. McCarthy for

helpful discussions and suggestions throughout this work.Thanks to John Hobbie for providing laboratory space and toJohn Helfrich for assistance in Woods Hole, Mass. during theinitial stages of this study. Thanks also to the MBL supply

department staff in Woods Hole for aid in obtaining Limnoriaspecimens. Appreciation is extended to Ed Seling for experttechnical assistance with the scanning electron microscope.Thanks also are given to H. Ducklow, D. Kirchman, R. Ka-puscinski, W. Kaplan, R. Sjoblad, S. Berk, B. Boyle, and J.Thiboutot for discussions, suggestions, and assistance duringthis study.

LITERATURE CITED

1. Alexander, M. 1971. Microbial ecology. John Wiley &Sons, Inc., New York.

2. Austin, B., S. Garges, B. Conrad, E. E. Harding, R.R. Colwell, U. Simidu, and N. Taga. 1979. Compar-ative study of the aerobic, heterotrophic bacterial floraof Chesapeake Bay and Tokyo Bay. Appl. Environ.Microbiol. 37(4):704-714.

3. Baumann, P., and L. Baumann. 1977. Biology of themarine enterobacteria: genera Beneckea and Photobac-terium. Annu. Rev. Microbiol. 31:39-61.

4. Baumann, L., P. Baumann, M. Mandell, and R. D.Allen. 1972. Taxonomy of aerobic marine eubacteria.J. Bacteriol. 110(1):402-429.

5. Berger, B., G. Thorington, and L. Margulis. 1979. Twoaeromonads: growth of symbionts from Hydra viridis.Curr. Microbiol. 3(1):5-10.

6. Boyle, P. J., and R. Mitchell. 1978. Absence of micro-organisms in crustacean digestive tracts. Science 200:1157-1159.

7. Boyle, P. J., and R. Mitchell. 1980. Interactions betweenmicroorganisms and wood-boring crustaceans, p. 179-186. In T. S. Oxley, D. Allsopp, and G. Becker (ed.),Biodeterioration: Proceedings of the Fourth Interna-tional Symposium, Berlin. Pitman Publishing Co., Lon-don.

8. Buchanan, R. E., and N. E. Gibbon (ed.). 1975. Bergey'smanual of determinative bacteriology. The Williams &Wilkins Co., Baltimore.

9. Carlucci, A. F., and S. L. Shimp. 1974. Isolation andgrowth of a marine bacterium in low concentration ofsubstrate, p. 363-367. In R. R. Colwell and R. Y. Morita(ed.), Effect of the ocean environment on microbialactivities. University Park Press, Baltimore.

10. Clarke, R. T. J. 1977. The gut and its microorganisms, p.35-72. In R. T. J. Clarke and T. Bauchop (ed.), Micro-bial ecology of the gut. Academic Press, Inc., New York.

11. Colwell, R. R., J. Kaper, and S. W. Joseph. 1977.Vibrio cholerae, Vibrio parahaemolyticus, and otherVibrios: occurrence and distribution in ChesapeakeBay. Science 198:394-396.

12. Colwell, R. R., and J. Liston. 1960. Microbiology ofshellfish. Bacteriological study of the natural flora ofPacific oysters (Crassostrea gigas). Appl. Microbiol. 8:104-109.

13. Colwell, R. R., and J. Liston. 1962. The natural bacterialflora of certain marine invertebrates. J. Insect Pathol.4:23-33.

14. Colwell, R. R., and R. Y. Morita (ed.). 1974. Effect ofthe ocean environment on microbial activities. Univer-sity Park Press, Baltimore.

15. Colwell, R. R., T. C. Wicks, and H. S. Tubiash. 1975.A comparative study of the bacterial flora of the he-molymph of Callinectes sapidus. Mar. Fish. Rev. 37(5-6) :29-33.

16. Colwell, R. R., and W. J. Wiebe. 1970. "Core" charac-teristics for use in classifying aerobic, heterotrophicbacteria by numerical taxonomy. Bull. Ga. Acad. Sci.28:165-185.

17. Davis, W. A., J. G. Kane, and V. F. Garagusi. 1978.Human Aeromonas infections: a review of the literatureand a case report of endocarditis. Medicine (Baltimore)57:267-277.

18. Dresel, E. I. B., and V. Moyle. 1950. Nitrogenous excre-


VOL. 42, 1981

tion of amphipods and isopods. J. Exp. Biol. 27:210-225.

19. Ducklow, H. W., and R. Mitchell. 1979. Bacterial pop-ulations and adaptations in the mucus layers on livingcorals. Limnol. Oceanogr. 24(4):715-725.

20. Fisher, E. S., L. S. Birnbaum, J. DePalma, J. S.Muraoka, H. Dear, and F. G. Wood. 1975. Surveyreport: Navy biological fouling and biodeterioration.Report NUC-TP 456. Naval Undersea Center, SanDiego, Calif.

21. Gibson, D. M., M. S. Hendrie, N. C. Houston, and G.Hobbs. 1977. The identification of some gram negativeheterotrophic aquatic bacteria. Aquat. Microbiol. 6:135-159.

22. Hamid, A., T. Sakata, and D. Kakimoto. 1976. Micro-flora in the alimentary tract of gray mullet. I. Isolationand identification of bacteria. Mem. Fac. Fish. Kago-shima Univ. 25(1):59-65.

23. Hamid, A., T. Sakata, and D. Kakimoto. 1977. Micro-flora in the alimentary tract of gray mullet. III. Studyon the characteristics of intestinal microflora in differ-ent conditions. Mem. Fac. Fish. Kagoshima Univ. 26:79-87.

24. Hamid, A., T. Sakata, and D. Kakimoto. 1978. Micro-flora in the alimentary tract of gray mullet. II. A com-parison of the mullet intestinal microflora in fresh andseawater. Bull. Jpn. Soc. Sci. Fish. 44(1):53-57.

25. Hartenstein, R. 1968. Nitrogen metabolism in the terres-trial isopod, Oniscus asellus. Am. Zool. 8:507-519.

26. Hobbie, J. E., R. J. Daley, and S. Jasper. 1977. Use ofNuclepore filters for counting bacteria by fluorescencemicroscopy. Appl. Environ. Microbiol. 33:1225-1228.

27. Jones, E. B. G., R. D. Turner, S. E. J. Furtado, andH. Kuhne. 1976. Marine biodeteriogenic organisms. I.Lignicolous fungi and bacteria and the wood-boringmollusca and crustacea. Int. Biodeterior. Bull. 12(4):120-134.

28. Kadota, H. 1959. Cellulose-decomposing bacteria in thesea, p. 332-341. In D. L. Ray (ed.), Marine boring andfouling organisms. University of Washington Press, Se-attle.

29. Kaneko, T., and R. R. Colwell. 1975. Adsorption ofVibrio parahaemolyticus onto chitin and copepods.Appl. Microbiol. 29(2):269-274.

30. Kaneko, T., R. R. Colwell, and F. Hamons. 1975.Bacteriological studies of Wicomico River soft-shellclam (Mya arenaria) mortalities. Chesapeake Sci.16(1) :3-13.

31. Kaper, J. B., H. Lockman, and R. R. Colwell. 1981.Aeromonas hydrophila: ecology and toxigenicity of iso-lates from an estuary. J. Appl. Bacteriol. 50:359-377.

32. Kaper, J., H. Lockman, R. R. Colwell, and S. W.Joseph. 1979. Ecology, serology, and enterotoxin pro-duction of Vibrio cholerae in Chesapeake Bay. Appl.Environ. Microbiol. 37(1):91-103.

33. Kirby, P. K., and R. D. Harbaugh. 1974. Diurnal pat-terns of ammonia release in marine and terrestrial iso-pods. Comp. Biochem. Physiol. 47(A):1313-1321.

34. Koneman, E. W., S. D. Allen, V. R. Dowell, Jr., andH. M. Sommers. 1979. Color atlas and textbook ofdiagnostic microbiology. J. B. Lippincott Co., Philadel-phia.

35. Kriss, A. E., I. E. Mishustina, N. Mitskevich, and E.V. Zemstova. 1967. Microbial populations of oceansand seas (translated by K. Syers). St. Martin's Press,New York.

36. Leibovitz, L. 1978. Bacteriologic studies of Long Islandshellfish hatcheries: an abstract. Mar. Fish. Rev. 40(10):


8.37. Liston, J. 1957. The occurrence and distribution of bac-

terial types on flatfish. J. Gen. Microbiol. 16:205-216.38. Lovelace, T. E., H. Tubiash, and R. R. Colwell. 1968.

Quantitative and qualitative commensal bacterial floraof Crassostrea virginica in Chesapeake Bay. Proc. Natl.Shellfish. Assoc. 58:82-87.

39. Merkel, J. R., and E. Traganza. 1958. Possible sym-biotic role of proteolytic and cellulolytic bacteria foundin the digestive system of a marine isopod. Bacteriol.Proc. 1958:53-54.

40. Merkel, J. R., E. D. Traganza, B. B. Mukherjee, T. B.Griffin, and J. M. Prescott. 1964. Proteolytic activityand general characteristics of a marine bacterium, Aero-monasproteolytica sp. n. J. Bacteriol. 87(5):1227-1233.

41. Murchelano, R. A., and J. L. Bishop. 1969. Bacterio-logical study of laboratory-reared juvenile Americanoysters (Crassostrea virginica). J. Invert. Pathol. 14:321-327.

42. Murchelano, R. A., and C. Brown. 1968. Bacteriologicalstudy of the natural flora of the Eastern oyster, Cras-sostrea virginica. J. Invert. Pathol. 11:519-520.

43. Oliver, A. C. 1962. An account of the biology of Limnoria.J. Inst. Wood Sci. 9:32-91.

44. Ramasamy, K., K. Prakasam, J. Bevers, and H. Ver-achtert. 1979. Production of bacterial proteins fromcellulosic materials. J. Appl. Bacteriol. 46:117-124.

45. Ray, D. L. 1959. Nutritional physiology of Limnoria, p.40-60. In D. L. Ray (ed.), Marine boring and foulingorganisms, University of Washington Press, Seattle.

46. Ray, D. L., and J. R. Julian. 1952. Occurrence of cellu-lase in Limnoria. Nature (London) 169:32.

47. Rheinheimer, G. 1971. Aquatic microbiology. John Wiley& Sons, Inc., New York.

48. Rippey, S. R., and V. J. Cabelli. 1980. Occurrence ofAeromonas hydrophila in limnetic environments: rela-tionship of the organism to trophic state. MicrobialEcol. 6(1):45-54.

49. Sieburth, J. M. 1979. Sea microbes. Oxford UniversityPress, New York.

50. Sleeter, T. D., P. J. Boyle, A. M. Cundell, and R.Mitchell. 1978. Relationships between marine micro-organisms and the wood-boring isopod Limnoria tri-punctata. Mar. Biol. 45:329-336.

51. Sokal, R. R., and F. J. Rohlf. 1969. Biometry, theprinciples and practices of statistics in biological re-search. W. H. Freeman and Co., San Francisco.

52. Unkles, S. E. 1977. Bacterial flora of the sea urchinEchinus esculentus. Appl. Environ. Microbiol. 34(4):347-350.

53. Vind, H. P., H. Hockman, and J. Muraoka. 1956. Theprotein requirements of Limnoria. Non-technical ab-stract no. 8. U.S. Naval Engineering Research andEvaluation Laboratory, Port Hueneme, Calif.

54. Wieser, W. 1966. Copper and role of isopods in degrada-tion of organic matter. Science 153:67-69.

55. Wieser, W., and G. Schweizer. 1970. A reexaminationof the excretion of nitrogen by terrestrial isopods. J.Exp. Biol. 52:267-274.

56. Wieser, W., G. Schweizer, and R. Hartenstein. 1969.Patterns in the release of gaseous ammonia by terres-trial isopods. Oecologia (Berlin) 3:390-400.

57. Yoshimizu, M., and T. Kimura. 1976. Study on theintestinal microflora of salmonids. Fish Pathol. 10(2):243-259.

58. Zachary, A., and R. R. Colwell. 1978. Gut-associatedmicroflora of Limnoria tripunctata in marine creosote-treated wood pilings. Nature (London) 282:716-717.

Documents

External Microflora of Marine Wood-Boring Isopod · Wood-boring isopods are ubiquitous in tem-perate and tropical marine habitats. They are responsible for substantial damage to equip-ment,